Write a script (or notebook) to gather info about notebooks in this repo

[agstephens]

In order to help users navigate their way around this repository of CEDA Jupyter Notebooks, we would like to have a top-level index file. This file, index.ipynb, should be auto-generated from a template/script/notebook. We should also have a non-interactive version of the page: index.md

It should provide include:

1. A brief intro about the repository
2. A basic search capability that matches against any software packages used in notebooks and datasets used.
3. A table of all the notebooks in the repo (with links to them) - and columns for datasets and keywords.
4. The state of the search should probably exist in a JSON/pickle file that lives in the repo.

@jhaigh0: Please create a prototype script that will:

• search the notebooks
• find all imported packages and record them as package_tags for that notebook
• find all things that look like CEDA datasets (e.g. /badc/<dataset>/ , /neodc/<dataset>/) - and record them as being dataset_tags for that notebook
• serialise the content to a JSON/YAML/pickle file
• build a basic search functionality in a notebook to allow search on that JSON/YAML/pickle file
• build a table of all the notebooks, including columns for Packages and Datasets
• consider whether there are other things that it is worth us tagging